I run a AMD 64 X2 Dual Core Processor 3800+. I have read and installed a optimizer for SETI, but i was wondering if anyone has heard of one for Prediction.... this project demands a lot of time. I was wondering what the average completion time is...

-edit: Entered right name of CPU

The program has been optimised in the compiler. Which isn\'t easy, as some options to speed it up cause the models to become unstable and crash rather quickly.

The program is over a million lines of Fortran, original written in 64 bit code for the Met Office\'s supercomputers. Optimising that to speed it up isn\'t a simple matter. It apparently took a couple of years just to port it to 32 bit desktop code that would run under Windows and still be stable.

On a 3.2GHz P4, running 24/7, with no other projects, and running little other software, it takes about 90 days to complete a model.

---

Backups: Here

Thank you that is exactly what i needed to know, But as with SETI, again, Does AMD run is slower than a normal Pentium equivalent?

AMD is quicker running the model than Pentium 4, but slower than the new Core Duo chips from Intel. Until the new Core microarchitecture was released, most of the top PCs in the project had AMD chips.

The X2 3800 overclocks well, you might be able to add 30-40%, but note that with a long duration project like this one, stability is more important than speed.

---

I'm a volunteer and my views are my own.
News and Announcements and FAQ

Donavan: I guess this phenomenon depends very much on the project. It is true that SETI favours Intels, especially the models with a large cache. But in Einstein, for example, any halfway decent AMD wipes the floor with its Intel equivalent (the Intel mobile architectures do okay, though, or better than ordinary Pentiums in any case) and Macs are worst off. It depends on the science app; certain code does better with certain CPU architectures. Therefore some of the optimized apps, but with those, you really have to know what you\'re doing or you\'re risking stability issues aswell.
FYI: I\'ve got a 3500+ and I\'m more than content with its CPDN performance, so, don\'t worry about that ;-)

Donavan: I guess this phenomenon depends very much on the project. It is true that SETI favours Intels, especially the models with a large cache. But in Einstein, for example, any halfway decent AMD wipes the floor with its Intel equivalent (the Intel mobile architectures do okay, though, or better than ordinary Pentiums in any case) and Macs are worst off. It depends on the science app; certain code does better with certain CPU architectures. Therefore some of the optimized apps, but with those, you really have to know what you\'re doing or you\'re risking stability issues aswell.
FYI: I\'ve got a 3500+ and I\'m more than content with its CPDN performance, so, don\'t worry about that ;-)

All my PC\'s but one are Intels and the other is an AMD Duel Core. From what members of the Team have stated and demonstrated AMD\'s are better at most of the projects with the Exceptions of the ones stated. The science is what I do the projects for but the credits keep us all interested.

---

Proud Founder and member of

Have a look at my WebCam